Comments on: Usage Data for X-Plane 10 & 11

By: Luke173

Luke173 — Wed, 22 Feb 2017 15:30:21 +0000

Could you also provide us with the airport usage data? This would sure be very interessting for Gateway Artists. Very interessing stats so far!

By: Peter Raslik

Peter Raslik — Tue, 21 Feb 2017 12:06:57 +0000

In reply to Tyler Young. That is true, if you can tell the proportion of noise data is marginal to useable data. I'm not saying that it isn't, just that the noise is generated by both developers and users, if not even much more by users.

By: WingsStayOn

WingsStayOn — Mon, 20 Feb 2017 21:42:26 +0000

Thanks for the usage data, it’s an interesting read. Although, as a scientist, I cannot consider valid any data that doesn’t give the sample size. In other words, I miss the “n” that’s so prominent (indeed, crucial) in any statistic. I understand that for commercial reasons you’re loathe to publish the exact numbers, but “hundreds of thousands of flights” just doesn’t cut it. What matters *a lot* is the balance of the XP10 versus the XP11 sample size, and from these data that’s impossible to gauge. I’d imagine that “many” flights in XP11 still is a lot less than “many” flights in XP10 – skewing the whole analysis. The specs & habits of a (relatively) small number of early adopters could wreck the whole image. In particular the hefty changes in the small percentages could well be entirely due to such effects (in biology I’d call them “founder effects”). I’ll out myself here and now as one of that small (but “very vocal” – yay !) group of Linux users and while we’re used to be squeezed aside by the bulk of much bigger numbers I especially object when that happens because of people’s blind faith in nicely coloured lines going up and down. There’s no *too* many statistics-savvy people out there, and without even the absolute basic denominator there’s no end to the conclusions that people can come to….

best regards,
WingsStayOn

By: Tyler Young

Tyler Young — Mon, 20 Feb 2017 14:35:10 +0000

In reply to Tyler Young. Another way of looking at this: random noise in the data set isn't an issue as long as it's dwarfed by the real data.

By: Tyler Young

Tyler Young — Mon, 20 Feb 2017 14:34:13 +0000

In reply to Bob Marsh. To the extent that developer use (or any non-flight activities in general) introduces noise into the data, we expect it to introduce noise *in proportion* to the real use cases they're testing for. So, for instance, if a developer opens the same work-in-progress airliner a hundred times, you would expect that the *real* users of that airliner would use it orders of magnitude more than that. (It would be very strange if that were not the case!)

By: Tyler Young

Tyler Young — Mon, 20 Feb 2017 14:30:47 +0000

In reply to Ulrich. It's not that we, as a company, are ignoring demo users... far from it! But we provide this data primarily for the sake of third-party scenery/aircraft/plugin developers, and for their purposes, demo users are really *not* their target market. (It would be very strange, for instance, for a demo user to buy a third-party add-on!)

By: Peter Raslik

Peter Raslik — Mon, 20 Feb 2017 12:17:34 +0000

In reply to Bob Marsh. Maybe flights shorter than 5-7 minutes should be discarded from the stats. Scenery testing, plugin testing, performance tuning, weather assessments and bugs checking are common scenarios which are usually performed under this time frame and may skew the results. On the other hand, simple circuit around the airport will probably take more than 7 minutes. And I would say focus should be on how simulator is used and not on how it is fiddled with. Or if it is important as well, divide the data into 2 groups - all flights and short flights excluded.

By: Ulrich

Ulrich — Mon, 20 Feb 2017 08:26:01 +0000

Why ignoring the Demo users? They are prospective customers (In my case I had v5, skipped versions 6 and 7, got v8 (was very much disappointed from the fact that 8.0 updated to 8.20 was different from an 8.20 being sold, making it impossible to use the scenery I had bought), got v9, skipped v10, now watching v11beta…
I also guess Windows 10 is forcing users to look for alternatives. Except from lacking speech output X-Plane 11 works amazingly well under Linux, even on an older OpenSUSE 13.2 (which is obsolete now).

By: Bob Marsh

Bob Marsh — Sun, 19 Feb 2017 04:40:26 +0000

If you can get to the pilot flight log in both X-Plane 10 and 11, and look at the number of flight hours, and aircraft there, and the flight activity, those records should be more enlightening than the data you are presently collecting. This is especially true because all developers are using X-Plane as a debug tool for Scenery and AC design. I suspect that your statistics are very hard to use as a gauge for comparison between user types, and true Simulator Flight usage. The real results, as you say, are hard to nail down for true Simulator Users actually flying their aircraft.

By: Elios

Elios — Sun, 19 Feb 2017 01:16:53 +0000

In reply to Ben Supnik. even the current Mac Pro which is now 4+ years old gets spanked by a budget windows box with a 150 buck GPU and with the Pro using both GPUs which it cant for Xplane again i think LR needs to take a hard look if its even worth supporting Metal or just move to DX11 with better dev tools and leave current OpenGL for people that MUST stay on Mac